General synthetic aperture radar (SAR) image automatic target recognition (ATR) methods perform well under standard operation conditions (SOCs). However, they are not effective in extended operation conditions (EOCs). To improve the robustness of the ATR system under various EOCs, an ATR method for SAR images based on the scattering parameter Gaussian mixture model (GMM) is proposed in this paper. First, an improved active contour model (ACM) is used for target–background segmentation, which is more robust against noise than the constant false alarm rate (CFAR) method. Then, as the extracted attributed scattering center (ASC) is sensitive to noise and resolution, the GMM is constructed using the extracted ASC set. Next, the weighted Gaussian quadratic form distance (WGQFD) is adopted to measure the similarity of GMMs for the recognition task, thereby avoiding false alarms and missed alarms caused by the varying number of scattering centers. Moreover, adaptive aspect–frame division is employed to reduce the number of templates and improve recognition efficiency. Finally, based on the public measured MSTAR dataset, different EOCs are constructed under noise, resolution change, model change, depression angle change, and occlusion of different proportions. The experimental results under different EOCs demonstrate that the proposed method exhibits excellent robustness while maintaining low computation time.

In electronic warfare environments, online anti jamming decision-making for frequency-agile radars must cope with unknown adversaries, limited interaction budgets, and non stationarity, where standard reinforcement learning can be sample inefficient and slow to adapt. We propose a Posterior Sampling for Reinforcement Learning (PSRL)–centric online decision frame work with dual-channel non-stationarity detection on rewards and transitions under a non-stationary Markov decision process (MDP) model. On the reward side, a Bernoulli-mean confidence-bound test is used to detect shifts in detection outcomes; on the transition side, a L1-deviation statistic is applied to empirical multinomial transition distributions. Detection thresholds are derived in closed form and explicitly tied to a target false-alarm probability, enabling principled parameter tuning. When either channel signals a change, the corresponding Bayesian posterior parameters are reset to discard stale information and accelerate adaptation to the new regime. Simulations across multiple stationary and non-stationary jamming scenarios evaluate convergence speed, recovery time, recovery rate, and cumulative return, and include sensitivity analyses over window length and false-alarm level to quantify the trade-off among detection sensitivity, delay, and false alarms. Results show that in stationary environments the detector remains inactive and preserves PSRL's sample efficiency, while in non-stationary environments the proposed method achieves faster recovery and improved steady state performance, yielding sustained gains in cumulative return.